Coherence in spoken discourse requires explicit connections between ideas, signposting of structure, and clear topic progression that accounts for the ephemeral nature of speech. Unlike written prose, spoken language cannot be re-read, so speakers must work harder to establish and maintain logical flow through repetition, explicit transitions, and thematic consistency. Incoherent speech confuses listeners and undermines message retention.
Prepare two versions of a 3-minute speech: one with minimal transitions and topic shifts, one with explicit signposting and transitions. Deliver both to different audiences and compare comprehension via a brief quiz.
From your work on cohesion and coherence, you know the difference: cohesion is the surface-level linking of sentences through devices like pronouns, connectives, and lexical repetition, while coherence is the deeper property of ideas hanging together into a unified, followable whole. In writing, both properties can be established once and held by the page — readers can scan backward, re-read a confusing sentence, or pause and return. In live speech, nothing holds. The moment you speak a word, it is gone, and the only record your audience has is what they managed to encode in working memory. This ephemerality is the foundational constraint that every spoken coherence strategy must address.
The practical consequence is that spoken coherence requires more explicit structure than written coherence, not less. The three-part formula — "tell them what you'll tell them, tell them, tell them what you told them" — is not cliché redundancy; it is an acknowledgment of working memory limits. Signposting announces the structure of what's coming so listeners build a schema before encountering the content: "I want to make three points about the budget — first, its assumptions; second, its risks; third, an alternative." Once that schema is established, each new idea slots into a pre-built container rather than floating free. Absent signposting, listeners spend cognitive resources building the schema mid-flight, leaving less capacity to process the content itself.
Transitions in speech must be more explicit than in writing because listeners cannot see paragraph breaks or heading changes. A new paragraph in an essay signals a topic shift visually; in speech, you must signal it verbally: "Now, moving from the evidence to what it implies for policy..." or even a deliberate pause followed by a change in vocal register. Internal summaries — brief restatements of the point just made before moving to the next — serve a repair function: they give listeners who drifted a moment to catch up, and they reinforce encoding for those who were following. Research on recall consistently shows that audiences remember signposted information better than equally important information that was not marked.
Thematic consistency is coherence at the macro level. A coherent speech returns to its central claim throughout — not by mechanical repetition, but by making each new point visibly connect to the core thread. The audience should be able to answer at any moment: "What is this speech about?" and "Where are we in the argument?" If those questions have unclear answers, coherence has broken down. One practical diagnostic: record a speech, then listen back and mark every moment where the throughline goes invisible. Those gaps are your revision targets. Incoherent speeches are rarely caused by bad ideas — they're caused by ideas that never got organized into a structure the speaker made audible.